Automated Seizure Detection, Quantification, Warning and Therapy Delivery Using the Slope of Heart Rate

Abstract

A system and method for detecting seizures in a patient are disclosed. The method includes measuring at least one signal associated with the patient over a first time period, and calculating a first slope having a rate of change in a first signal of the at least one signal over the first time period. The first signal is a heart rate. The method also includes determining a difference between the first slope and a reference slope, as well as detecting an epileptic seizure based at least upon the difference between the first slope and the reference slope. Furthermore, the method includes quantifying the epileptic seizure by calculating a severity of the seizure. The severity is the product of a seizure duration and a seizure intensity.

What is claimed is: 
     
         1 . A medical device for detecting seizures in a patient, comprising:
 a cardiac sensor configured to measure a heart rate of the patient over at least a first time period equivalent to three heart beats of the patient;   a controller communicatively coupled to the cardiac sensor and configured to:   receive the heart rate from the cardiac sensor;   calculate a first slope, the first slope being at least an estimate of a rate of change in the heart rate over the first time period;   determine a difference between the first slope and a reference slope, the reference slope being a reference value for changes in R-R interval durations during non-ictal conditions; and   detect an epileptic seizure based at least upon the difference between the first slope and the reference slope;   wherein the first slope is calculated by comparing a time interval between a first R wave peak and a second R wave peak with a time interval between the second R wave peak and a third R wave peak.   
     
     
         2 . The medical device of  claim 1 , wherein the reference slope is based upon one or more heart rate measurements performed under inter-ictal circumstances including at least one of physical activity, resting, high ambient temperature, daytime, and nighttime. 
     
     
         3 . The medical device of  claim 1 , wherein the controller is further configured to receive at least one autonomic signal in addition to the heart rate and to utilize the at least one autonomic signal, along with the first slope, in detecting the epileptic seizure, wherein each autonomic signal is one of cardiac, vascular, respiratory, and dermal. 
     
     
         4 . The medical device of  claim 3 , wherein the epileptic seizure is detected based upon a cluster analysis performed by the controller in a state space having a dimension for each of the at least one autonomic signals and a dimension for the first slope, wherein the cluster analysis employs a centroid defined, at least in part, using the reference slope. 
     
     
         5 . The medical device of  claim 1 , wherein the detection of the epileptic seizure occurs in real-time. 
     
     
         6 . The medical device of  claim 1 , wherein the controller is further configured to:
 calculate a seizure duration and a seizure intensity, and then quantify the epileptic seizure by calculating a severity of the epileptic seizure;   wherein the severity is the product of the seizure duration and the seizure intensity;   wherein the seizure duration is the amount of time the difference between the first slope and the reference slope spent outside a threshold value; and   wherein the seizure intensity is a maximum value of the first slope during the epileptic seizure.   
     
     
         7 . The medical device of  claim 6 , wherein the controller is further configured to classify the detected epileptic seizure in response to determining that at least one of the seizure intensity and the seizure duration are above the 75th percentile or below the 25th percentile of values for representative seizures. 
     
     
         8 . The medical device of  claim 1 , wherein the cardiac sensor is implemented as part of a wearable computing device, and wherein the controller is communicatively coupled to the cardiac sensor through the wearable computing device. 
     
     
         9 . The medical device of  claim 8 , wherein the controller is also part of the wearable computing device. 
     
     
         10 . The medical device of  claim 9 , wherein the wearable computing device is a smart watch. 
     
     
         11 . A method for detecting seizures in a patient, comprising:
 receiving at a controller, a heart rate of the patient measured by a cardiac sensor over at least a first time period equivalent to three heart beats of the patient,   calculating a first slope with the controller, the first slope being at least an estimate of a rate of change in the heart rate over the first time period;   determining, with the controller, a difference between the first slope and a reference slope, the reference slope being a reference value for changes in R-R interval durations during non-ictal conditions; and   detecting, with the controller, an epileptic seizure based at least upon the difference between the first slope and the reference slope;   wherein the first slope is calculated by comparing a time interval between a first R wave peak and a second R wave peak with a time interval between the second R wave peak and a third R wave peak.   
     
     
         12 . The method of  claim 11 , wherein the reference slope is based upon one or more heart rate measurements performed under inter-ictal circumstances including at least one of physical activity, resting, high ambient temperature, daytime, and nighttime. 
     
     
         13 . The method of  claim 11 , further comprising:
 receiving, at the controller, an autonomic signal in addition to the heart rate;   wherein detecting the epileptic seizure further comprises utilizing the autonomic signal, along with the first slope;   wherein the autonomic signal is one of cardiac, vascular, respiratory, and dermal.   
     
     
         14 . The method of  claim 13 , wherein detecting the epileptic seizure further comprises performing a cluster analysis in a state space having a dimension for each of the at least one autonomic signals and a dimension for the first slope, the cluster analysis using a centroid defined, at least in part, by the reference slope. 
     
     
         15 . The method of  claim 11 , wherein the detection of the epileptic seizure occurs in real-time. 
     
     
         16 . The method of  claim 11 , further comprising:
 calculating a seizure duration and a seizure intensity;   quantifying the epileptic seizure by calculating a severity of the epileptic seizure;   wherein the severity is the product of the seizure duration and the seizure intensity;   wherein the seizure duration is the amount of time the difference between the first slope and the reference slope spent outside a threshold value; and   wherein the seizure intensity is a maximum value of the first slope during the epileptic seizure.   
     
     
         17 . The method of  claim 16 , further comprising classifying, with the controller, the detected epileptic seizure in response to determining that at least one of the seizure intensity and the seizure duration are above the 75th percentile or below the 25th percentile of values for representative seizures. 
     
     
         18 . The method of  claim 11 , wherein the cardiac sensor is part of a wearable computing device, and wherein the controller is communicatively coupled to the cardiac sensor through the wearable computing device. 
     
     
         19 . The method of  claim 18 , wherein the controller is also part of the wearable computing device. 
     
     
         20 . The method of  claim 19 , wherein the wearable computing device is a smart watch.